Chris Anderson, editor of Wired, deeply brilliant man, and a friend, has made the best case for the Long Tail; and, indeed, there’s a wealth of anecdote and an intellectual plausibility to support it. Back in the early days of the Internet, when its root server stood unguarded on a shelf in a Virginia basement,  I marveled at the existence of a company - in Indiana, if memory serves - from which you could buy every conceivable type of hubcap. Thanks to the Internet (it wasn’t the Web yet), weak signals of demand could be discerned and aggregated - and a store created to serve the three people in the whole wide world who needed a hubcap for a 1954 Hudson.

Well, no. Professor Anita Elberse of Harvard Business School cut the Long Tail up and turned it into oxtail soup in a Harvard Business Review article in July 2008. It turns out, if you look at the numbers, that the Web is really an incredibly powerful amplifier of the feast-or-famine, bestseller-or-flop tendencies that characterize all global media. Just recently  Charles M Blow of the New York Times took another whack at the long tail;  in a column analyzing the sad state of music sales, he cited data from PRS for Music, a UK association of songwriters and composers: they show that, of  the 13 million songs for sale online, only 3 million had any sales last year at all; and 52,000 songs snagged 80 per cent of the revenue, a ratio that makes Pareto’s 80-20 rule look like a great deal for the little guy.

When you think about it, Long Tail theory is based on the premise that there exists a vast market of people with esoteric taste who never before had the means to satisfy it - millions upon millions of geeky grad students whose pockets bulge with unspent cash. Those customers exist - just not in big numbers. They can support a hubcap business, but not an automobile industry. In the real world of the Web, the rich get richer, and the niches get nichier. As Ann Winblad of the Hummer Winblad VC firm once told me, “If there’s a long tail, I want the fat part.”

The dream of ad-based revenue models is doomed to disappoint, too, for three reasons. By now even the most gung-ho Web startups recognize that there will never be more than a handful of megasites - the Googles and YouTubes that attracts tens or hundreds of millions of visitors and can therefore command ad revenue on the basis of audience alone. That hasn’t stopped people from imagining an ad-sales version of the Long Tail, in which ever-more-infinite customer knowledge produces the ability to define and pitch to ever-more-infinitesimal market segment: If I create a destination website for professional rodent removers, then whoever builds a better mousetrap will have to beat a path to my door. The fallacy there is the one that afflicts Long Tail thinking generally: while there are, indeed, micro-mini markets, they command micro-mini revenue; most people still want what most people want. Marketers’ hoary lament is “I know that half of my advertising spending  is wasted; I just don’t know which half.” The hope of  ad-based sites is that they can solve that problem. But suppose they did? The obvious response of marketers would be to cut their spending in half.

Dreams of advertising riches are naïve for a second reason: They throw an infinite supply of advertising inventory against a finite supply of advertising spending. There is - let’s face it - only so much ad budget to go around: X per cent or Y per cent of GDP, depending on whether you’re in a consumer economy like the U.S. or EU or an emerging one like India or China; A per cent or B per cent of revenue, depending on whether you’re selling consumer goods or gas turbines. When the number of advertising outlets expands (in the US, from three television networks in the 1960s to hundreds today, for example), each gets a smaller piece of pie. It’s worse on the Web, where there is no theoretical limit to the amount of advertising a site can create. When demand is constant and supply increases, prices fall: you learned that, recall, on the first day of your first economics class.

The third nail in the coffin of the eyeballs-to-adverts delusion is what my Booz & Company colleagues brilliantly call “private label media”. These days companies can literally disinter-MEDIA-te by investing in their own media - professional, popular Web sites that they build and operate themselves. Why advertise on Epicurious (the foodie web site operated by Conde Nast) when you can build BettyCrocker.com and attract 8 million unique visitors a month?

Private label media is the latest, and most powerful, manifestation of a long-term trend, whereby advertising has taken a smaller and smaller portion of all marketing spending - a stepchild of marketing. According to analyst Jack Myers, 70 per cent of marketing communications spending went into advertising in 1963; today, however, only 30 per cent does - the rest going into merchandising, trade promotion - and now, private-label media. Sure, marketers still use advertising to attract customers - but increasingly they are trying to move them into their private-label media, and keep them there. Check out babycenter.com, for example, run by Johnson & Johnson: when expectant parents register there, they are invited to enter their baby’s due date. On the site, they get great content - produced by an editorial staff every bit as good as that of magazines for parents. And J&J, knowing the baby’s birthday, can estimate when the child will be ready for a first pair of shoes, for toilet training, for a tricycle - even for acne medication. If you were a rational marketing executive, where would you put your next dollar of spending? 

And if you were a rational entrepreneur, how would you design your business plan?

Thomas A. Stewart is the Chief Marketing and Knowledge Officer of Booz & Company. Formerly the Editor and Managing Director of Harvard Business Review, he is the author of Intellectual Capital: The New Wealth of Organizations and The Wealth of Knowledge: Intellectual Capital and the 21st-Century Organization.